*Copyright @ www.mycsg.in;

Create a sample dataset

This dataset contains symbols digit strings lowercase letters and uppercase letters
It helps us study how SAS sorts character values by default and how the sort sequence can be modified
Notice that some values look numeric such as `10` and `101` but they are stored as character values in the variable `var`

data series01;
   length var group $10;
   var="%"; group="Symbol"; output;
   var="$"; group="Symbol"; output;
   var="1"; group="Digits"; output;
   var="2"; group="Digits"; output;
   var="10"; group="Digits"; output;
   var="101"; group="Digits"; output;
   var="21"; group="Digits"; output;
   var="11"; group="Digits"; output;
   var="a"; group="Lowcase"; output;
   var="ab"; group="Lowcase"; output;
   var="A"; group="Uppercase"; output;
   var="AB"; group="Uppercase"; output;
   var="b"; group="Lowcase"; output;
   var="B"; group="Uppercase"; output;
run;

Inspect `series01` before sorting and notice that the observations are in creation order not sorted order
The variable `group` is included only to help you identify what kind of character value each row contains

Understand the default sort sequence of character variables in SAS

With the default sort sequence SAS orders character values according to internal collating rules
In a simple English environment the order is usually symbols then digit strings then uppercase letters then lowercase letters
Character digits are sorted character by character not by numeric value
That means `101` can come before `11` because SAS compares from left to right as text

Create a dataset named series01_sort01 by sorting series01 in ascending order of var

`proc sort` sorts the observations based on the character variable `var`
Review the sorted output carefully and notice the relative position of symbols, digits, uppercase values, and lowercase values
Also compare the order of `1`, `10`, `101`, `11`, `2`, and `21` to see character-based ordering in action

proc sort data=series01 out=series01_sort01;
   by var;
run;

The result demonstrates the default collating behavior for character data
This example is important because many programmers initially expect character digits to be sorted numerically but that is not the default

Modify the sorting sequence of character variables using the SORTSEQ option

The `sortseq=` option lets us change how SAS compares character values during sorting
The `linguistic` sequence provides more natural language aware sorting behavior than the default binary style collation

Group lowercase and uppercase letters together

The `linguistic` option treats uppercase and lowercase forms more naturally
This makes values such as `A` and `a` sort closer together than they do under the default sequence

proc sort data=series01 out=series01_sort02 sortseq=linguistic;
   by var;
run;

Compare `series01_sort02` with the default sorted dataset and observe how letter case is handled differently

Group lowercase and uppercase letters together and give preference to lowercase letters

The `case_first=lower` suboption tells SAS to place lowercase values before uppercase values when the letters are otherwise equivalent
This can be useful when a business rule or output requirement expects lowercase values to appear first

proc sort data=series01 out=series01_sort03 sortseq=linguistic(case_first=lower);
   by var;
run;

Review the order of pairs such as `a` and `A` or `ab` and `AB` and confirm that lowercase values are preferred

Group lowercase and uppercase letters together and give preference to uppercase letters

The `case_first=upper` suboption does the reverse and places uppercase values first when the base letters match
This demonstrates that sort order can be tuned to meet display or reporting conventions

proc sort data=series01 out=series01_sort04 sortseq=linguistic(case_first=upper);
   by var;
run;

Compare the position of uppercase and lowercase values against the previous example and confirm that uppercase values now appear first

Sort character digits based on their numeric value

The `numeric_collation=on` suboption tells SAS to compare embedded digit strings by numeric value rather than pure character order
This is useful when character variables store codes like `1` `2` `10` and `101` and you want them sorted numerically

proc sort data=series01 out=series01_sort05 sortseq=linguistic(numeric_collation=on);
   by var;
run;

Check the order of the digit values in `series01_sort05` and compare it with `series01_sort01`
You should now see a more natural numeric progression for the digit strings

Key points to remember

Character values are not always sorted in the way users intuitively expect
Default sorting compares character data using collating rules rather than numeric meaning
`sortseq=linguistic` can provide more natural handling of letter case
`case_first=` controls whether lowercase or uppercase values come first
`numeric_collation=on` helps character digit strings sort by numeric value

*Copyright @ www.mycsg.in;

Create a sample dataset

SAS Log

Dataset View

Understand the default sort sequence of character variables in SAS

Create a dataset named series01_sort01 by sorting series01 in ascending order of var

SAS Log

Dataset View

Modify the sorting sequence of character variables using the SORTSEQ option

Group lowercase and uppercase letters together

SAS Log

Dataset View

Group lowercase and uppercase letters together and give preference to lowercase letters

SAS Log

Dataset View

Group lowercase and uppercase letters together and give preference to uppercase letters

SAS Log

Dataset View

Sort character digits based on their numeric value

SAS Log

Dataset View

Key points to remember